Before It Was Obvious

You're sitting in front of your computer and you want to build a startup. You've seen Cursor sell to Elon for $60B . Maybe the previous generation was Mark Zuckerberg or Evan Spiegel. You look at these founders and compare yourself to them. They don't seem that much smarter than you. Their resumes aren't better than yours.

So naturally you ask yourself: Why couldn't I do the same? This is where most founders begin. And this is also where most founders get stuck. They see AI. They see crypto. They see thousands of startups already funded. Every category looks crowded. Every obvious idea has already been built. They conclude there are no opportunities left. They close their laptop and move on. And that's where a large percentage of startups die. Not because the founders weren't capable, but because they assumed the game was already over. Let's use Cursor as an example. Not every path is a straight shot. Cursor has been chewing glass since 2022. This was before ChatGPT. There wasn't a playbook. There wasn't an obvious market. There was only a belief that AI would fundamentally change knowledge work. To stay grounded they focused on three things. First, they chose an area that genuinely excited them: AI. Second, they were customers of their own product. Third, they focused relentlessly on power users. Because if you can win power users, everyone else becomes much easier. And to be quite frank, this story isn't unique to Cursor.

Stripe started when online payments already looked solved, but the founders believed developers would increasingly become decision makers inside companies and that whoever won developers would eventually win the internet. They had experienced the pain firsthand, and while PayPal had already proven online payments worked, Stripe saw an opportunity to build the developer first version of that future.

Figma spent years building before the market was ready because they believed the future of design wasn't better design tools, but collaborative design where everyone worked in the same file. Google Docs had already demonstrated the power of real time collaboration for docs. Figma extended that insight to design.

Shopify began as an attempt to sell snowboards online because the founders believed millions of small businesses wanted to own their customers, brand, and destiny rather than depend on marketplaces. Amazon had proven centralized e-commerce could work. Shopify bet founders would eventually want to own it themselves. Different products. Same pattern.

Each founder started with a non consensus belief about where the world was heading and then spent years building before that future became obvious to everyone else. The luck was being attached to powerful tailwinds.

In Stripe's case, it was the belief that more and more commerce would move online. In Figma's case, it was the belief that software would become cloud first and collaborative by default. In Shopify's case, it was the belief that the internet would empower millions of entrepreneurs to build independent businesses.

Cursor follows a similar path. The company was built around the belief that AI would fundamentally change knowledge work and that software engineers would be among the first power users to adopt it. The product may look obvious today, but when they started there was no clear playbook. Only conviction. Different products. Different markets. The same underlying pattern. Identify a secular shift early, find the wedge others are missing, and spend years executing before the rest of the market catches up. There will always be a yin to the yang. PayPal led to Stripe. Adobe led to Figma. Amazon led to Shopify.

The first generation proves the market exists. The second generation reimagines it around a new insight, new technology, or changing customer behavior. The important question for founders is figuring out where they are in the cycle. If you're early, like Coinbase or Cursor, the opportunity is often making the tech usable for power users. Coinbase didn't invent crypto. It simply made buying and holding Bitcoin dramatically easier than managing wallets yourself or wiring money to Mt. Gox. Cursor didn't invent AI coding. It realized that autocomplete wasn't the end state and that developers wanted an AI native way to build software. But if you're building in the middle or later stages of a tech shift, the opportunity often looks different. The infrastructure already exists. The market has already been proven. Your job is no longer to prove that the technology works, but to find the yin to the existing yang the missing insight that the first generation overlooked. That's where many of the biggest companies are born.

So now you've identified where you are in the technological shift. You have a few ideas, you're ready to build, and then you realize something uncomfortable: you don't actually have many insights. You don't deeply understand the market, the customers, or even the products. And that's normal.

This is where you have to roll up your sleeves and start building your network, insights, and reputation. Thankfully, we live in a post X world where this is easier than ever. You can build an audience, meet customers, interact with power users, and learn directly from the people shaping the market.

The first thing I would do is use every product in the space. If you're building in a category and you're not a power user of the products that define it, it's very difficult to develop unique insights about where the market is heading. Map every product in the ecosystem. Become a power user of each one. Talk to the people who love them, the people who hate them, and the people who have switched away from them. Understand why they stay, why they leave, and what they wish existed that doesn't today.

Eventually you'll discover that most markets aren't won because incumbents are stupid. They're won because incumbents become successful. As companies grow, they naturally move further away from individual users. Feedback loops slow down, edge cases get ignored, and new power users emerge that don't fit the existing product. This is where new founders find opportunity.

The goal isn't to come up with an idea in isolation. The goal is to immerse yourself so deeply in a market that the missing pieces become obvious. Once you've done that long enough, you'll stop looking for ideas and start noticing them everywhere. And this is where you want to be. Eventually you'll realize there are more opportunities than you could ever build. Now comes the hard part: choosing one.

Once you've identified what you believe is the right idea, the next question is simple: is this a 10x improvement or a hair on fire problem? If the answer is no, don't bother. People rarely switch for a marginal improvement. They switch when something is dramatically better or when the pain is so severe that they need a solution immediately.

The easiest way to find a hair on fire problem is to find people already building workarounds. Spreadsheets, WhatsApp groups, manual processes, copying data between systems these are all signals. The best founders are searching for pain because when the pain is large enough, customers will pull the product out of your hands. When the pain is small, no amount of marketing, growth hacks, or clever positioning will save you. Now you've identified the idea, the pain, and you're building an MVP.

With Claude, Codex, and AI tooling, it's never been easier to build products. Ironically, that's become its own trap. I found myself adding feature after feature simply because I could. The product slowly turned into a Frankenstein like product. Every feature sounded reasonable in isolation, but together they made the product worse.

Eventually I went back to first principles. The most important question isn't what features should I build. It's why would someone switch from what they're already using?

Every great startup has an answer to that question. Cursor could have built another coding plugin. Instead, they forked VS Code. Developers already loved the editor, understood how it worked, and had it embedded in their daily workflow. Cursor wasn't asking users to learn something entirely new. It was asking them to keep doing what they already loved, just with AI built directly into the experience.

The best startups rarely force users to learn entirely new behaviors. Instead, they find familiar workflows, remove friction, and make them dramatically better.

As founders, we obsess over what we're building. Customers obsess over what they have to give up. The lower the switching cost and the higher the value created, the faster adoption happens. That's why the best MVPs aren't feature rich. They're painfully focused on giving customers a compelling reason to switch.

At this point you've identified the pain, built the MVP, and hopefully given customers a compelling reason to switch. Now comes the part most founders underestimate: distribution.

I see founders obsess over product for months and spend five minutes thinking about how users will discover it. The truth is that distribution is often the moat.

Airbnb didn't win because its website was better. The founders knocked on doors, took photos of apartments themselves, and manually onboarded hosts city by city. Stripe recruited developers one at a time. Coinbase lived in Bitcoin forums long before crypto became mainstream.

Cursor is another great example. The team posted on Hacker News six different times. Most of the posts went nowhere. They DMed thousands of developers, listened obsessively to feedback, and earned users one at a time. Today everyone acts like Cursor was inevitable. For years they were doing things that didn't scale.

Founders love talking about product-market fit, but before product-market fit comes distribution-market fit. Where do your customers spend time? Who do they trust? How do they discover new products? The best founders don't just build products. They build distribution engines. Because the market can't love a product it never sees. The last phase of all of this is grit, resilience, and not giving up.

Unfortunately, I can't teach you this. Nobody can. It has to be experienced. Cursor is again a great example. They spent years building before the market was ready. They posted repeatedly, DMed thousands of users, and most people ignored them. In hindsight it looks obvious. In real time it looked anything but.

The same pattern shows up everywhere. Airbnb's founders were rejected repeatedly and at one point funded the company by selling cereal boxes. Nvidia survived multiple near death experiences before becoming one of the most valuable companies in the world. Rain, a startup from one of our batches, launched in the aftermath of FTX when most people thought crypto was dead. While others were leaving the industry, they kept building. Years later they raised over $100M at a $2B valuation.

The lesson isn't that these founders were smarter than you. It's that they stayed in the game long enough for their insights to compound.

And so I've laid out the framework for you.

Find the technological shift. Develop a unique insight. Become obsessed with the market. Talk to customers. Find a hair on fire problem. Build the simplest possible wedge. Earn distribution. And most importantly, don't quit when things get hard.

That's it.

There is no secret. Most people won't do these things consistently enough for long enough. The few that do end up building the companies the next generation of founders will study.

The world is yours.

Go build.